Metrics on N and the distribution of sequences

A set A belongs to \text{D} if and only if the limit \lim_{N \to \infty} E_N(\mathcal{X}_A) := d(A) exists. \text{D} is a q - algebra and d is density on \text{D}.

A set A \subset {\mathbb N} belongs to \text{D} if and only if \overline{d}(A)+\overline{d}({\mathbb N} \setminus A) \le 1.

A sequence \{v(n)\}, v(n) \in [0,1] is uniformly distributed modulo 1 if and only if for each h \in {\mathbb N} we have

If the sequences \{v_1(n)\}, \dots, \{v_k(n)\} of elements of some interval [a,b], a<b have continuous asymptotic distribution function then these sequences are statisticaly independent if and only if

for x_1, \dots, x_k real numbers, where F_j is asymptotic distribution function of the sequence \{v_j(n)\}, j=1, \dots, k.

If the sequences \{v_1(n)\}, \dots, \{v_k(n)\} satisfy the conditions of 4 then the sequence \{v_1(n)+\dots + v_k(n)\} has continuous asymptotic distribution function.

Each borel probability measure on metric space is regular.

The system of the sets of P - continuity \mathcal{B}_c(M) is an algebra of sets.

For each \alpha \in {\mathbb M} and r>0 the set B(\alpha, r') belongs to \mathcal{B}_c(M) for r'\in [0,r] with except of countable subset of [0,r].

For each \alpha \in {\mathbb M} and r >0 such increasing sequence of r_n < r that r_n \to r, n \to \infty that the open balls B(\alpha, r_n) are the set of P - continuity.

An open base of the metric space ({\mathbb M}, \mathfrak{d}) exists that it contains only the open balls which are the sets of P - continuity.

The system \text{D}_\nu is an algebra of sets and \nu = \nu^\ast |_{\text{D}_\nu} is density on \text{D}_\nu.

A set S \subset {\mathbb N} is a \nu^\ast - measurable set if and only if its closure cl(S) is the set of P - continuity.

Each sequence uniformly continuous with respect to \mathfrak{d} has weak \nu - distribution function.

For each x - point of continuity of F_0 we have F_0(x)= P(\tilde{v}\le x).

If \{v(n)\} is a bounded sequence uniformly continuous with respect to \mathfrak{d} having a continuous \nu - distribution function F then for every a<b we have

Let \{v(n)\} be a sequence of elements of some interval [a,b], a<b having absolutely continuous \nu - distribution function F. If f:[a,b] \to [a,b] is such continuous function that for each x \in [a,b] we have \lambda(f^{-1}(\{x\}))=0, then the sequence \{f(v(n))\} continuous \nu - distribution function.

Let \{v_1(n)\}, \dots, \{v_k(n)\} be uniformly continuous sequences with respect to \mathfrak{d}, having continuous \nu - distribution functions. If these sequences are \nu - independent then the random variables \tilde{v}_1, \dots, \tilde{v}_k are independent.

For each S \subset {\mathbb N} the inequality \overline{d}(S) \le \nu^\ast(S) holds.

If S is \nu^\ast - measurable set then S \in \text{D} and \nu(S)=d(S).

Each naturally continuous sequence has weak asymptotic distribution function. If this sequence is dense in some interval then this weak asymptotic distribution function is increasing on this interval.

If \{v(n)\} is naturally a continuous sequence of interval [a,b], a< b, with continuous asymptotic distribution F, then for subinterval [c,d] \subset [a,b] the set v^{-1}([c,d]) belongs to \text{D}_\nu and

If a sequence \{v(n)\} is naturally continuous and \tilde{v}(\alpha)=0 for \alpha \in S where P(S)=1 then v(n)=0 for each n \in {\mathbb N}.

If \{v(n)\} is a sequence is naturally continuous, and for each \varepsilon >0 we have P(|\tilde{v}|> \varepsilon)=0 then \tilde{v}(\alpha)=0 for each \alpha \in {\mathbb M}.

If \{v(n)\} is naturally continuous then D^2(v)=0 if and only if this sequence is constant.

If the sequences \{u(n)\}, \{v(n)\} are jointly naturally continuous then \rho(v,u)=1 if and only u(n)=av(n)+b for n \in {\mathbb N}.

Let \{u(n)\}, \{v(n)\} be jointly naturally continuous sequences of elements of [0,1] . Then

If and only if E(v)=E(u), E(u^2)=E(v^2), E(uv)=E(v^2).

If \{v(n)\}, \{u(n)\} are sequences of elements from given interval [a,b], a<b, jointly naturally continuous , having the same weak asymptotic distribution function and E(uv)=E(v^2) then u(n)=v(n) for n \in {\mathbb N}.

A metric \mathfrak{d} is natural if and only if for for every sequence uniformly continuous with respect the metric \mathfrak{d} i) the limit \lim_{N \to \infty}E_N(v)=E(v) exists and ii) E(v)>0 whenever \{v(n)\} is non negative sequence strongly positive on some closed ball with center from {\mathbb N}.

\mathfrak{d} is metric.

Let S be a set dense in real line. A sequence \{v(n)\} have continuous asymptotic distribution function if and only if i) v^{-1}((-\infty, x)) belongs to \text{D} for x \in S and ii) for each real number a we have \lim_{\eta \to 0^+} \overline{d}(v^{-1}((a-\eta, a+\eta)))=0.

Let f be such continuous real function defined on interval [a,b] that such division a=a_0 < a_1 < \dots < a_m=b exists that the function f is strictly monotone on [a_i, a_{i+1}], i=0, \dots, m. If \{v(n)\} is a function having continuous asymptotic distribution function, then the sequence \{f(v(n)\} have continuous asymptotic distribution function also.

The sequences \{v_k(n)\}, k=1,2,3, \dots are jointly naturally continuous.

The random variables \tilde{v}_1, \dots, \tilde{v}_k, \dots are independent.

If f is such real continuous function defined on [a,b] that the sequences \{f(v_i(n)\} have continuous asymptotic distribution function, then for almost all \alpha \in {\mathbb M} we have

Suppose that f is real valued continuous monotone function defined on [a,b]. For almost all \alpha \in {\mathbb M} exists non increasing sequence of \varepsilon_n >0 ; n \in {\mathbb N} such for each sequence of natural numbers \{k_n\} where k_n \in B(\alpha, \varepsilon_n) we have

For almost all \alpha \in {\mathbb M} such decreasing sequence \varepsilon_n >0 exists that for each sequence of natural numbers \{k_n\} such that k_n \in B(\alpha,\varepsilon_n) the sequence \{v_n(k_n)\} is uniformly distributed modulo 1.